Diagnosis and management of chronic myeloid leukemia

A survey of American and European practice patterns

Authors


Abstract

BACKGROUND.

The success of imatinib therapy for chronic myeloid leukemia (CML) has brought new challenges; these include optimizing disease monitoring, imatinib resistance, and use of novel, more potent tyrosine kinase inhibitors. Thus, there is a need to establish new best practices for CML management in the post-imatinib era.

METHODS.

An internet-based questionnaire, consisting of 26 multiple choice questions, was developed to assess hematologists' and oncologists' self-reported treatment strategies for CML.

RESULTS.

Between November 2005 and January 2006, 956 eligible physicians responded to the survey; 727 from the US and 229 from Europe. Most US respondents (60%) practiced in the community setting, whereas European respondents were primarily academic (44%) and hospital-based (40%). Physicians' responses were generally in line with current recommendations, although differences were identified. Confusion existed among respondents over optimal timing of treatment decisions, with a notable proportion of physicians focusing on a single timepoint rather than consistent monitoring, as currently advocated. Some respondents were unaware of new molecular monitoring techniques, when to monitor for BCR-ABL mutations (and the impact on treatment decisions), and the benefit of new tyrosine kinase inhibitors.

CONCLUSIONS.

Responses to the survey suggest that treatment practices in some areas of CML management are not in line with current recommendations. Identified areas of need should be targeted in future educational activities for the CML community. Cancer 2007. © 2007 American Cancer Society.

The treatment of chronic myeloid leukemia (CML) has changed dramatically over the last decade. As recently as 1997 the standard therapy for patients with newly diagnosed chronic phase CML not eligible for allogeneic stem cell transplant (SCT) was interferon alpha (IFN-α), alone or with low-dose cytosine arabinoside (araC).1 The realization that the initiating event in CML was the creation of the BCR-ABL fusion protein led to considerable effort directed toward the development of agents that would target this constitutively activated tyrosine kinase.2 These efforts culminated in the development of imatinib, a selective BCR-ABL inhibitor that altered the treatment paradigm for CML, shifting the focus away from allogeneic SCT as frontline therapy by 2002.3

This rapid change in clinical practice was driven largely by the results of the IRIS (International Randomized Study of Interferon versus STI571) trial, which compared imatinib with IFN-α plus low-dose araC in patients with newly diagnosed CML.4 Follow-up of the IRIS trial is now in excess of 5 years5 and it has become clear that the prognostic risk factors that applied to traditional therapies, such as IFN-α, are not as applicable to imatinib treatment. Whereas a patient's Sokal score (based on age, spleen size, and platelet and peripheral blood blast count) does correlate with their likelihood of achieving a complete cytogenetic response (CCyR) with imatinib, and the associated survival benefits,6 the depth of response to imatinib is a better indicator of long-term outcome that overrides pretherapy variables. In the IRIS study, 97% of patients with a major molecular response (MMR) to imatinib at 12 months were progression-free at 54 months follow-up, irrespective of their Sokal risk group at baseline. The prognosis for those patients who achieved CCyR without MMR at 12 months was also good. However, among patients who failed to achieve a CCyR at 12 months, 28% had progressed at 54 months follow-up.7, 8

The prognostic value of MR seen in the IRIS study has meant that the monitoring of minimal residual disease has become increasingly important. Rising levels of BCR-ABL transcripts are an early indication of loss of response, and thus the need to reevaluate treatment. Regular molecular monitoring of individual patients is clearly warranted but there is much variation between laboratories in testing methods and in the levels of access to such tests. These variations have prompted a recent drive to standardize molecular and disease monitoring in CML.9

In addition to changes in disease monitoring, resistance to imatinib has led to the development of several new treatment options that overcome the various mechanisms of imatinib resistance. The availability of new, more potent kinase inhibitors such as dasatinib (SPRYCEL, formerly BMS-354825) and others currently in clinical development (eg, nilotinib, SKI-606, INNO-406) add to the pressing need for optimizing treatment decisions, and for these decisions to be linked to appropriate disease monitoring. Dasatinib was recently approved for the treatment of all phases of CML and Philadelphia-positive acute lymphoblastic leukemia (Ph-positive ALL) with resistance or intolerance to prior therapy, including imatinib.

Given the rapid rate of change in CML management at present, a greater understanding of current practices on matters such as new treatment options, reimbursement issues surrounding these treatment options, disease monitoring, and access to proper monitoring tools is needed. Identifying physicians' preferences in CML monitoring and response criteria, and the rationale for treatment choices, may identify additional areas of need, and so aid in focusing educational activities to the CML community. This article reports the results of a survey conducted to evaluate general trends of practice in CML treatment and monitoring among hematologists and oncologists within the US and in Europe. The findings of the survey are discussed in the context of current published guidelines and clinical trial outcomes.

MATERIALS AND METHODS

An internet-based questionnaire was developed to assess physicians' self-reported management of CML. The survey was developed in collaboration between the coauthors (H.K., J.C.), the Network for Medical Communication and Research (NMCR), and Oncology Network Europe (ONE); the 26 questions used and the multiple choice responses are shown in Table 1. A physician was considered eligible to participate if he/she had treated ≥4 patients with CML outside the context of a clinical trial in the 2 years preceding the survey. This stipulation ensured that the survey captured practice patterns outside the directive of a clinical trial.

Table 1. Questions and Answers in the Online Physician Survey
No.QuestionResponse options
  1. CML indicates chronic myeloid leukemia; SCT, stem cell transplantation; FISH, fluorescence in situ hybridization; IFN, interferon; PCR, polymerase chain reaction; QPCR, Quantitative PCR; CHR, complete hematologic response; CyR, cytogenetic response; McyR, major CyR; CcyR, complete CyR; Ph, Philadelphia chromosome.

1In the past 2 years have you treated four or more patients with CML outside the context of a clinical trial?Yes; No
2How would you characterize your medical practice?Academic; Hospital; Community; Combination
3How many patients with CML are currently under your direct management?1–5; 6–10; 11–15; 16–20; >20
4Which of the following best describes the composition of your practice?Hematologic malignancies only; Both solid and hematologic; Mostly solid tumors
5Which of the following is indicative of the number of years of clinical experience you have had since completing your training?<1 year; 1–5 years; 6–10 years; 11–15 years; 16–20 years; >20 years
6In a newly diagnosed 40-year old patient with chronic phase CML who has a matched related sibling, which of the following would you recommend?Imatinib 400 mg/day; Imatinib 600–800 mg/day; Immediate allogeneic SCT; Other
7In a newly diagnosed 40-year old patient with chronic phase CML who has an unrelated matched donor, which of the following would you recommend?Imatinib 400 mg/day; Imatinib 600–800 mg/day; Immediate allogeneic SCT; Other
8When a patient of yours receives a preliminary diagnosis of CML, where do you obtain confirmatory testing by cytogenetics and/or FISH?Local hospital; Nearest teaching hospital lab; Genzyme; Quest; Lab Corp.; Other commercial lab
9When a new patient of yours receives a preliminary diagnosis of CML, where do you obtain confirmatory testing by qRT-PCR and or BCR-ABL kinase domain mutation?Local hospital; Nearest teaching hospital lab; Genzyme; Quest; Lab Corp.; Other commercial lab
10When evaluating a patient with a white cell count of 40,000 and a morphologic picture consistent with CML, which of the following initial diagnostic tests would you perform? Please choose all that applyCytogenetics; FISH; Quantitative PCR; Other
11At what point do you consider it most critical to evaluate response in order to decide whether to continue or change therapy?1 month; 2 months; 3 months; 6 months; 12 months; 3, 6, 12 then every 3–6 months; Other
12In addition to peripheral blood counts, which of the following do you use in monitoring response to imatinib therapy? Please choose all that applyCytogenetics; FISH; Quantitative PCR; BCR-ABL kinase domain mutation analysis; Other
13AHow frequently do you typically repeat cytogenetic analysis studies for monitoring response to imatinib therapy?Every 3 months; Every 6 months; Yearly; Not repeated; Other
13BHow frequently do you typically repeat qRT-PCR for BCR-ABL mRNA studies for monitoring response to imatinib therapy?Every 3 months; Every 6 months; Yearly; Don't utilize qRT-PCR
13CHow frequently do you typically repeat BCR-ABL kinase domain mutation studies for monitoring response to imatinib therapy?Only for patients who fail to show CyR to imatinib; Only for patients who have a 2-fold rise in BCR-ABL; Never ordered the test; Unavailable; Not familiar with this test
14In a 70-year old woman with Philadelphia chromosome-positive CML on imatinib 400 mg orally daily, which of the following strategies would you use for monitoring response to imatinib treatment? Please choose all that applyBlood counts; Cytogenetics; FISH; Quantitative PCR; Other
15A 50-year old patient with Philadelphia chromosome-positive CML is on imatinib daily and has a matched related sibling donor. For this patient, which of the following monitoring strategies would you initiate? Choose all that apply.Blood counts; Cytogenetics; FISH; Quantitative PCR; Other
16A 45-year old patient with CML on imatinib mesylate therapy has a matched related sibling BM donor. Which of the following would cause you to decide that the patient had experienced a suboptimal response to imatinib and should proceed with an alternative treatment?No CHR at 3 months; Ph + 100% at 6 months; Ph+ >35% at 12 months; Any Ph+ at 12 months; Any Ph+ at 18–24 months; QPCR + in CCyR; QPCR+ 0.4% in CCyR; QPCR+ 1% in CCyR at 24 months; 2-fold rise in QPCR in CCyR; ≥1 log rise in QPCR in CCyR; Loss of CHR; Increase in Ph+ by 30%; <3 log reduction in QPCR at 12 months; None of the above
17Which of the following do you feel best indicate “effective” imatinib therapy when treating newly diagnosed chronic phase CML patients?CCyR in 3 months; CCyR in 6 months; CCyR in 12 months; MR in 6 months; MR in 12 months; Both CCyR and MR in 12 months
18How do you usually assess imatinib-associated toxicity in your patients?Frequent physician visits early in course of therapy; Frequent nurse visits early in course of therapy; Toxicity questionnaire completed by patients; Telephone query performed by nurse; Other
19Which of the following imatinib toxicities have you encountered in your patients with CML? Please choose all that applyNausea and vomiting; Diarrhea; Rash; Neutropenia; Thrombocytopenia; Anemia; Muscle cramps; Liver dysfunction; Fluid retention; Periorbital edema; Pleural effusion; Pericardial effusion or pericarditis; Weight gain; Congestive heart failure; Fatigue; Bone aches; Other (please specify)
20Which of these toxicities have been severe enough to cause you to interrupt (I), dose reduce (R), or discontinue (D) imatinib therapy?Nausea and vomiting; Diarrhea; Rash; Neutropenia; Thrombocytopenia; Anemia; Muscle cramps; Liver dysfunction; Fluid retention; Periorbital edema; Pleural effusion; Pericardial effusion or pericarditis; Weight gain; Congestive heart failure; Fatigue; Bone aches; Other (please specify)
21Which of the following would cause you to decide that response to imatinib was suboptimal and to consider a change in therapy? Please choose all that applyFailure to achieve CHR at 3 months; Failure to achieve any CyR at 6 months; Failure to achieve MCyR at 1 year; Failure to achieve CCyR at 18–24 months; Failure to achieve QPCR negativity at 12–24 months
22What do you generally recommend for patients with a suboptimal response to 400 mg of imatinib daily?Increase imatinib to 600 mg/day, evaluate in 3–6 months; Increase imatinib to 800 mg/day, evaluate in 3–6 months; Allogeneic SCT for HLA matched patients; Allogeneic SCT if response suboptimal after 3–6 months of imatinib 600–800 mg/day
23For patients with suboptimal response to imatinib do you assess for BCR-ABL kinase domain mutations?Yes; No
24How do you manage CML patients who have suboptimal response to dose-escalated imatinib and who do not have an HLA matched bone marrow donor?Continue imatinib and add investigational agents on a clinical trial (eg, tipifarnib or lornafarnib); Continue imatinib and add IFN, AraC, or hydroxyurea; Discontinue imatinib and begin IFN ± AraC; Discontinue imatinib and begin dasatinib or nilotinib; Other
25For a 65-year old patient with CML who has failed imatinib (using your standard criteria), and who has a matched related sibling, which of the following treatment options would you choose next?Immediate allogeneic SCT; Dasatinib or nilotinib; Low dose IFN, AraC, hydroxyurea, or combination of these; Other investigational strategy
26For a 65-year old patient with CML who has failed imatinib (using your standard criteria), who has a matched unrelated donor, which of the following treatment options would you choose next?Immediate allogeneic SCT; Dasatinib or nilotinib; Low dose IFN, AraC, hydroxyurea, or combination of these; Other investigational strategy

Of approximately 15,000 oncologists and hematologists invited to participate in the survey, 956 eligible physicians responded between November 2005 and January 2006 (6.4% response rate), 727 from the US and 229 from Europe; 267 labeled themselves as ‘academic’ and 689 as ‘nonacademic’ (hospital, community, and combination practice settings). Among US-based responders the academic label was 23% vs 44% among European responders. Responses were collated and analyzed by the authors of this article. The contents of this study represent key conclusions drawn from these data.

RESULTS

Physician/Practice Characteristics

The first section of the questionnaire focused on the characteristics of physicians and their practices. Respondents were grouped according to practice setting (academic, hospital, community, or combination). Most US respondents (60%) practiced in the community setting. In contrast, European respondents were primarily based in academic (44%) and hospital (40%) settings, with only 7% in community practice. These figures may not necessarily be indicative of standard practice patterns in the US and Europe, but rather reflective of the respondents to this survey. Additionally, the understanding of the terms ‘hospital-based’ and ‘community-based’ may have differed between US and European physicians and, indeed, between individual physicians in the same geographical region. Nevertheless, variations in response between the US and Europe to the remainder of the questionnaire may be a consequence of these differing practice settings rather than of the geographic region. Country-specific differences in practice patterns within Europe are known to exist, but were not examined in this survey.

There were marked differences in practice composition between US and European respondents. Most US respondents (75%) manage both solid and hematologic cancers, whereas European physicians were more specialized, with 54% managing hematologic malignancies only. This result may be a reflection of practice setting: 45% of academic respondents manage only hematologic malignancies, whereas 73% of nonacademic respondents manage both solid and hematologic malignancies. Patient numbers managed by US and European physicians were comparable; however, three times as many European respondents currently had >20 CML patients under their care. This is likely due to the predominance of hematologists among European respondents vs their US counterparts. Survey participants did vary in terms of number of years of clinical experience after completion of training: one-third of US respondents and 13% of European respondents had <5 years of experience. More than one-third of European respondents had 6–10 years of practice experience, whereas approximately 20% of US respondents had >20 years of experience vs 16% of their European counterparts. This question did not take into account the number of patients treated by physicians during the indicated timeframe; therefore, the duration of clinical experience quoted should not be confused with the extent of clinical experience.

Practice Patterns by Patient Type

In the next section, respondents were presented with several hypothetical clinical scenarios and asked to select their favored treatment option.

Case example

In a newly diagnosed 40-year-old patient with chronic phase CML and a matched related sibling donor, most respondents (62%, US; 79%, Europe) chose standard-dose imatinib (400 mg) over allogeneic SCT (13%, US; 10%, Europe) or high-dose imatinib. Details of the patient's Sokal score were not given, which could have influenced a decision for transplantation. Higher-dose imatinib (600–800 mg) was more frequently chosen as an option in the US vs Europe (24%, US; 9%, Europe), a distinction which may be driven by differences in indications for imatinib and the fact that in Europe such off-label use of imatinib would not be reimbursed.

Case example

For an older (65-year-old) patient who had failed imatinib therapy and had a matched related sibling, approximately 50% of all respondents chose the new tyrosine kinase inhibitors dasatinib or nilotinib as their next option. This was compared with 26% and 19% of US and European respondents, respectively, who opted for immediate allogeneic SCT. The trend for dasatinib or nilotinib first rather than SCT may be appropriate, given this patient's age. Interestingly, a substantial proportion of respondents (13% in the US and 26% in Europe) stated that they would choose IFN-α, low-dose araC, or hydroxyurea alone or in combination. However, until recently the use of novel tyrosine kinase inhibitors was confined to clinical trials and it is not clear which, or how many, respondents had access to them.

Case example

For the same 65-year-old patient, but with a matched unrelated donor, the proportion of respondents who would initiate dasatinib or nilotinib treatment increased from 49% to 55% in the US and from 49% to 57% in Europe, as one might expect. There was a concomitant decrease in the proportion of respondents who would choose immediate SCT. Yet, the preference of a minority of physicians (18% in the US and 25% in Europe) for pre-imatinib therapies such as IFN-α, araC, and hydroxyurea remained.

Disease Monitoring

Molecular monitoring of BCR-ABL transcript levels by quantitative real-time polymerase chain reaction (qRT-PCR) is increasingly used to assess treatment response in CML patients.9, 10 Molecular monitoring is particularly relevant when residual levels of leukemic cells are reduced to levels below detection by cytogenetic analysis. It is recommended by some CML experts that BCR-ABL transcript levels in patients treated with tyrosine kinase inhibitors are measured every 3 months from treatment initiation.3, 10 To gain an understanding of physicians' preferences in disease monitoring and response, and the rationale and triggers for their treatment choices, a number of questions in this area were included in the survey.

When questioned about the location of testing the vast majority (93%) of European respondents relied on either their local hospital laboratory (50%) or nearest teaching hospital laboratory (43%) for confirmatory cytogenetic and/or fluorescence in situ hybridization (FISH) studies, with extremely limited use of commercial laboratories. In contrast, half of all US respondents relied on commercial laboratories for these tests. A similar pattern was observed for qRT-PCR and BCR-ABL kinase domain mutation testing. This difference in Europe/US practices may be because commercial laboratories are more common in the US vs Europe, likely a reflection of the lower level of private healthcare utilization in Europe (although private laboratories are beginning to emerge in Europe). Alternatively, the disparity may be a reflection of academic vs community practice preference. When grouped by practice setting, most academic respondents indicated that they perform cytogenetics/FISH testing at either their local hospital laboratory (63%) or a nearby teaching hospital laboratory (26%), whereas nonacademic respondents tend to utilize more commercial laboratory testing. Another explanation may be the possible existence of affiliations between US insurance companies and specific private laboratories.

The survey respondents were also presented with several hypothetical patient cases, and questioned with regard to which diagnostic test(s) they would perform.

Case example

In a patient with a white cell count of 40,000/μL and a morphologic picture consistent with CML, 37% of European respondents stated that they would perform cytogenetics, FISH, and qRT-PCR as initial diagnostic tests (Fig. 1A). Nearly one-third (31%) of US respondents would perform these 3 tests in combination, but a similar proportion (33%) would utilize cytogenetics and FISH only. Over half of all US respondents and one-third in Europe would not implement qRT-PCR at diagnosis, seemingly at odds with the emphasis on molecular monitoring based on studies showing the prognostic significance of qRT-PCR data.10–13 Because both US and European studies support the use of qRT-PCR, the disparity between the US and European respondents may be a reflection of the greater proportion of pure hematologists (vs hematologist/oncologists) in the European vs US group in this survey. Alternatively, the US physicians may recognize the value of qRT-PCR monitoring at later timepoints, but are not aware of its value in the initial monitoring phase.

Figure 1.

Disease monitoring preferences of European and US physicians as determined by responses to questions (A) 10, (B) 14, and (C) 15 in the survey. B, blood counts; C, cytogenetics; F, fluorescence in situ hybridization; N, none of the above; NR, no response; O, other; Q, quantitative polymerase chain reaction monitoring.

Case example

In the case of a 70-year-old woman with Ph-positive CML on imatinib 400 mg/day, more European respondents (16%) would monitor imatinib response with blood counts, cytogenetics, and qRT-PCR, than with any other option offered in this survey (Fig. 1B). This omission of FISH analysis is aligned with the lower use of the technique in Europe in general, and the tendency to omit FISH analysis in elderly patients, as in this hypothetical setting. Among the US respondents there was no discernable trend with respect to preference for treatment monitoring. This may reflect the more frequent reliance on initial (first year) FISH rather than qRT-PCR monitoring in the US.

Case example

In a younger (50-year-old) Ph-positive CML patient with a matched sibling donor and treated with imatinib daily, nearly one-third of US respondents indicated that they would monitor with blood counts, cytogenetics, FISH, and qRT-PCR. More than half of European respondents would either use: blood counts, cytogenetics, FISH, and qRT-PCR (27%); or blood counts, cytogenetics, and qRT-PCR (24%) (Fig. 1C). These results indicate that only half of all respondents would monitor the disease of this hypothetical patient in line with recently published recommendations.3

Direct comparison between responses for the 50- and 70-year-old hypothetical patients discussed above indicates that, both in the US and Europe, there is a trend for more frequent monitoring in younger patients, especially those in whom allogeneic SCT is an option. Given that treatment options have been more limited for older patients, there is a tendency to reduce monitoring that may not drive treatment decisions. It will be interesting to see how the intensity of disease monitoring in older patients, and in general, changes as new CML treatment options become available.

Frequency of Disease Monitoring in CML and Influence on Treatment Decisions

Identification of the Ph chromosome by cytogenetic analysis remains the ‘gold standard’ for diagnosis of CML. After treatment initiation, guidelines recommend that cytogenetic analysis be repeated at 6- to 12-month intervals until the level of Ph-positive metaphases falls below 10%, when qRT-PCR should be implemented to monitor residual disease.14, 15 Disease monitoring by FISH is acceptable in cases where bone marrow cannot be obtained or examined successfully with conventional cytogenetics. Questions relating to frequency of disease monitoring were included in the survey to determine current practices.

Overall, 31% and 39% of US and European respondents, respectively, thought that patient response should be evaluated at 3, 6, and 12 months as a basis for treatment decisions, and then every 3–6 months, in line with current guidelines. When grouped by practice setting (academic vs nonacademic), the split of responses was similar. Somewhat surprisingly, one-third of US and one-quarter of European respondents felt that it was most critical to evaluate response at 3 months. Although the value of achieving treatment milestones at 3 months has been documented, changes in treatment based on this timepoint alone might be considered premature. This may suggest the need to educate physicians on the importance of continued disease monitoring in the first year of treatment and beyond, as recommended by the recent publication of the European LeukemiaNet expert panel.3

Most US respondents stated that they repeat cytogenetic analysis for monitoring response to imatinib every 3 months (34%) or every 6 months (33%), and nearly half (46%) of European respondents perform cytogenetic analysis every 6 months, whereas another 29% assess cytogenetics every 3 months. A notable proportion of respondents (19% in the US and 15% in Europe) reported that repeating cytogenetic analysis just once annually was sufficient. This question did not specify whether this was monitoring in the first or subsequent years of treatment. That aside, these data reflect the need to highlight the importance of continued bone marrow evaluation, not only for monitoring CyR, but also for assessing bone marrow morphology in the context of dose changes.

When queried on the frequency with which they would repeat qRT-PCR monitoring, 39% of US respondents stated they would do so every 3 months and 38% every 6 months. A higher proportion (53%) of European respondents monitor imatinib response by qRT-PCR every 3 months, whereas 31% do so every 6 months. When grouped by practice setting, 49% of academic respondents use qRT-PCR every 3 months, compared with 39% of nonacademic respondents. These data likely reflect the strong lead taken by European academic laboratories in molecular testing and may suggest a need to increase awareness of the importance of molecular monitoring disease by qRT-PCR among nonacademics. To aid this, more data may be required to confirm the value of molecular monitoring in a patient in CCyR.

More than 60% of US respondents were either not familiar with testing for BCR-ABL kinase domain mutations or had never ordered the test, with only 29% having done so. In contrast, almost 50% of European respondents had utilized BCR-ABL kinase domain mutation analysis, despite the fact that 26% of European physicians had no access to the test compared with 10% of US physicians. When grouped by practice setting, 60% of nonacademic respondents were either not familiar with BCR-ABL kinase domain mutation testing or had never ordered the test, compared with only 25% of academic respondents. Importantly, 43% of academic respondents would only order mutational analysis under specific circumstances, which were entirely appropriate. These data are reflective of a greater awareness in Europe of the importance of BCR-ABL mutations, and suggest there may be a need for greater education among US physicians on when to monitor for BCR-ABL mutations and, importantly, how to translate this information into treatment decisions. Although not analyzed in this survey, it is unclear whether most physicians, either in the US or Europe, understand how to interpret mutational data; this is likely because until very recently3 no recommendations existed on when to intervene or how to change treatment based on BCR-ABL mutational analysis. Excessive mutational testing may result in more patients being treated with SCT, with the associated significant morbidity and mortality risks, rather than other treatment modalities. However, the clinical significance of mutational testing and results in imatinib-sensitive patients are still unclear, even in the research setting, and requires further study.

Management of Suboptimal Response to Standard Dose Imatinib

The LeukemiaNet expert panel recently defined suboptimal response to imatinib for patients in chronic phase CML as a state where the patient may still have a substantial benefit from continuing imatinib, but that the long-term outcome of the treatment would not likely be as favorable.3 Specifically, these circumstances would be: less than a complete hematologic response (CHR) at 3 months; less than partial CyR (PCyR) (Ph-positive metaphases >35%) at 6 months; less than CCyR at 12 months; less than MMR (reduction of BCR-ABL/control gene ratio to ≤0.10% of a standard baseline value) at 18 months; the presence of additional chromosomal abnormalities in Ph-positive cells; a loss of MMR; or BCR-ABL mutations. Baccarani et al.3 propose that in the event of suboptimal response the imatinib dose should be increased to 600–800 mg/day and allogeneic SCT may be offered to eligible patients.

In line with these recommendations, more than three-quarters of US respondents would recommend increasing the imatinib dose to 600 mg/day (49%) or 800 mg/day (27%) for patients with a suboptimal response to 400 mg/day imatinib. In contrast, only 35% and 18% of European respondents would increase the imatinib dose to 600 and 800 mg/day, respectively. Almost twice as many European as US respondents (38% vs 18%) would increase the imatinib dose with the intention of advancing to SCT if the response remained suboptimal at 3–6 months.

When asked if they employed testing for BCR-ABL kinase domain mutations in patients with suboptimal imatinib response, most respondents (73%) in the US did not, whereas 50% in Europe did. This difference was outside of the academic vs nonacademic split, and highlights the need for more studies/education regarding the significance of BCR-ABL mutational analysis in this context.

When questioned on how they would manage suboptimal response in patients with CML and failure to respond to dose-escalated imatinib without a matched donor, approximately 50% of US and European respondents would discontinue imatinib therapy and enter patients in a clinical trial with dasatinib or nilotinib. This demonstrates a high level of awareness among both European and US physicians of the new kinase inhibitors currently available or in development for the treatment of CML. There was a trend for continued imatinib therapy with the addition of hydroxyurea, araC, or IFN-α (27% in Europe, 18% in US). A further 8% of US and 12% of European respondents stated that they would discontinue imatinib therapy and, rather than enrolling patients in a clinical trial with either dasatinib or nilotinib, would initiate IFN-α therapy, alone or with araC.

Defining Failure to Respond to Standard Dose Imatinib

Failure to respond to imatinib is defined as a clinical situation where continuing imatinib in any dose is no longer appropriate, and when alternative therapies should be considered.3 The parameters that indicate failure include: no hematologic response at 3 months; less than CHR at 6 months; less than a PCyR at 12 months; less than CCyR at 18 months; and loss of CHR, CCyR, or detection of a BCR-ABL kinase domain mutation with a low level of sensitivity to imatinib at any time.

The respondents in this survey were asked to define what they felt constituted failure to respond to imatinib in the context of a hypothetical patient.

Case example

When presented with a 45-year-old CML patient with a matched donor, there was little consensus among respondents on the definition of a failure of response to imatinib, and when to initiate alternative treatment. The most frequent responses are shown in Table 2.

Table 2. Definition of Failure to Respond to Imatinib According to Survey Respondents
Definition of failure of response% US respondents% European respondents
  1. CHR indicates complete hematologic response; CyR, cytogenetic response; McyR, major CyR; CcyR, complete CyR; Ph, Philadelphia chromosome.

No CHR at 3 mo1112
No CyR at 6 mo1211
Loss of CHR1212
No MCyR at 1 yr1010
No CCyR at 1 yr107
No CCyR at 18–24 mo88
30% increase in Ph-positive metaphases1011

Interestingly, when respondents were asked about what constitutes failure to respond, dictating a change in therapy in general, rather than related to a specific patient, both US and European physicians were aligned with current recommendations.3 Specifically, most respondents considered failure to respond as a lack of either a CyR at 6 months or a major CyR (MCyR) at 1 year. Among European respondents, 13% stated that they would declare failure to respond if the patient failed to achieve CHR after 3 months, CyR after 6 months, MCyR after 1 year, and CCyR after 18–24 months. Another 9% reported the same along with failure to achieve qRT-PCR negativity after 12–24 months. A total of 27% and 18% of US and European respondents, respectively, indicated that failure to achieve any CyR after 6 months (13%, US; 9%, Europe) or MCyR after 1 year (14%, US; 9%, Europe) should be considered failure to respond, and warrant consideration of alternative therapy. These responses are partly reflective of the new recommendations for defining imatinib failure as discussed above. It will be interesting to observe how these views evolve with increasing awareness of the new guidelines.

Defining an Effective Response to Imatinib

When questioned on what constitutes ‘effective’ imatinib therapy in newly diagnosed chronic phase CML, 30% and 38% of US and European respondents, respectively, defined effective response by both a CCyR and MR at 1 year; 30% and 25% of US and European respondents, respectively, stated that MR at 6 months constituted ‘effective’ imatinib therapy. The former result is reflective of experience with >5 years follow-up of the IRIS study, but the latter is puzzling given the lack of evidence for this milestone. The recent recommendations from Baccarani et al.3 support the use of hematologic and cytogenetic response parameters measured within the first year, but suggest that treatment decisions based on MR alone at 3 or 6 months are premature. The extent to which a physician's opinion on what constitutes effective imatinib therapy affects their treatment decisions is unclear.

Imatinib Toxicity

One of the advantages of molecular targeted therapy with tyrosine kinase inhibitors is the comparatively low rates of associated toxicity. Data from the IRIS study has shown that imatinib is generally well tolerated, with rates of intolerance significantly lower than with IFN-α/araC combination therapy.4 Nevertheless, a minority of imatinib-intolerant patients (3%–5%) experience toxicity so severe that it warrants treatment discontinuation.16

Most respondents in this survey (90%, US; 97%, Europe) assess imatinib-associated toxicity with frequent visits early in therapy. Similar incidences of imatinib toxicities were reported in both the US and Europe. When questioned on specific toxicities that had led to dose interruption, dose reduction, or imatinib discontinuation, responses (summarized in Table 3) were as one might expect, and comparable between the US and Europe. With the arrival of new kinase inhibitors, these responses are likely to change. For example, a physician who currently prescribes steroids to a patient with imatinib-associated rash in the future may discontinue imatinib in favor of dasatinib or nilotinib.

Table 3. Percentage of Physicians Who Would Discontinue, Interrupt, or Dose-Reduce Imatinib if Indicated Adverse Events Were Severe
Adverse event% Discontinue, US/Europe% Interrupt, US/Europe% Dose reduced, US/Europe
Nausea and vomiting4/1250/3645/52
Diarrhea5/542/4553/51
Rash13/2146/4041/39
Neutropenia6/1141/3853/51
Thrombocytopenia8/1039/4054/49
Anemia4/634/3662/58
Muscle cramps6/1143/3651/53
Liver dysfunction23/2244/4134/36
Fluid retention8/837/3055/61
Periorbital edema7/837/2056/73
Pleural effusions17/2035/3747/43
Pericardial effusion or pericarditis33/2232/4234/36
Weight gain4/934/2461/67
Congestive heart failure35/3029/4036/30
Fatigue6/638/1857/76
Bone aches3/540/2657/69

DISCUSSION

In the rapidly shifting landscape of CML therapy, an understanding of physicians' treatment preferences is of key importance in identifying areas for improvement in care and to inform future educational activities. Generally, the practice patterns of respondents to this survey were aligned with current guidelines and published clinical trials. However, there were some notable findings.

From the responses to questions on treatment choice, it is clear that SCT is becoming increasingly less common. Respondents consistently chose imatinib as the first-line treatment of newly diagnosed CML, despite the option of a matched related sibling donor; a trend in agreement with data showing improved survival rates and quality of life with imatinib vs SCT.17 Use of SCT may even be reduced to third- or fourth-line treatment with the emergence of high-dose imatinib/new tyrosine kinase inhibitor strategies. The survey indicated that the high-dose imatinib regimen was more common in the US vs Europe. To date, comparison of high-dose imatinib (800 mg/day) vs standard dose (400 mg/day) demonstrates that the higher dose provides increased rates of CCyR and earlier MR at the expense of greater myelosuppression; the long-term clinical significance of these differences remains to be demonstrated.18 Therefore, until further clinical data are available the practice patterns in Europe in this respect may be more appropriate than in the US.

Pre-imatinib treatment strategies such as hydroxyurea, araC, or IFN-α remain in use by some physicians as salvage treatment after imatinib failure and unsuitability of SCT, despite the emergence of newer tyrosine kinase inhibitors. Nevertheless, approximately half of all respondents chose dasatinib or nilotinib as second-line treatment. This figure is likely to rise as access to these drugs extends beyond clinical trials. Dasatinib was recently approved by the FDA for treatment of all phases of CML with resistance or intolerance to prior therapy, including imatinib, after successful Phase II trials conducted in over 900 patients.19–23 Initial Phase II trial results for nilotinib are also promising.24–26

It is clear from this survey that there is some confusion among physicians over optimal timing of treatment decisions: one-third of European and one-quarter of US respondents consider the 3-month timepoint to be the most critical for evaluating responses to therapy and making treatment decisions. Long-term follow-up of the IRIS trial shows that a CCyR to imatinib is an important response-related prognostic factor. Failure to achieve any CyR at 3 months does not necessarily signal failure of imatinib therapy, because patients still have a 50% chance of achieving CCyR at 2 years.27 Hematologic responses at 3 months do correlate with long-term outcome and should be considered when making treatment decisions; patients with no hematologic response at 3 months can be considered to have failed therapy, and patients who do not achieve a CHR at 3 months can be considered to have a suboptimal response. Overall, physicians should be mindful of the new recommendations for treatment, which advocate consistent evaluation of a range of responses 3–18 months postinitiation of treatment,3 rather than focusing on a single response criterion at a single timepoint.

Cytogenetic analysis is a major component of effective CML disease monitoring, but the benefit of molecular monitoring by qRT-PCR has increased greatly in the era of tyrosine kinase inhibitor therapy, where patients regularly achieve disease levels lower than can be detected accurately by cytogenetics. BCR-ABL transcript levels measured by qRT-PCR early in therapy can predict subsequent response and the probability of acquired resistance to imatinib.9, 10 Patients who achieve an MMR at 12 months have a very low probability of disease progression in the subsequent year.28 Evidence from this survey suggests that European laboratories use molecular monitoring with qRT-PCR more often, possibly because of the high proportion of academics among European responders and the fact that the test is well established in European academic laboratories. Major improvements in assay techniques have advanced PCR from a purely qualitative test with significant variability to a real-time quantitative assay with reproducible results. However, many physicians may not fully appreciate the difference between the qualitative and quantitative techniques and this may represent an area of educational need among physicians treating CML. Indeed, this confusion may have influenced responders when answering questions on the use of PCR in this survey.

According to the recent recommendations, if qRT-PCR reveals increasing BCR-ABL levels, patients should undergo analysis for BCR-ABL kinase domain mutations. A number of BCR-ABL mutants are known to be highly refractory to imatinib therapy, and their emergence should trigger a change in treatment strategy.3 Patients with any one of a further set of mutants conferring low levels of resistance to imatinib should be monitored closely. Most US respondents were either not familiar with tests for BCR-ABL kinase domain mutations or had never ordered the test, in contrast to their European counterparts. There is also greater awareness among European respondents of the specific circumstances in which it is appropriate to carry out the mutational analysis.3 These findings highlight the need to educate the CML community on the importance of BCR-ABL mutational analysis in certain situations, and the manner in which the data should be interpreted.

In summary, the results of this survey provide an intriguing snapshot of current treatment practices at this crucial point in the evolution of CML management. Given the availability of new tyrosine kinase inhibitors, and the dramatic change in our understanding of the value of disease monitoring based on clinical experience with imatinib, it will be of interest to see how the changing treatment paradigm affects practice patterns. To this end, a repeat of this survey in the future would be of interest to identify changing trends in treatment patterns.

Acknowledgements

Hagop Kantarjian and Jorge Cortes received research grants from Novartis and Bristol-Myers Squibb not related to this study. Andreas Hochhaus received research grants from Novartis, Bristol-Myers Squibb, and Wyeth. Michele Baccarani has received consultancies and honoraria from several companies, including Bristol-Myers Squibb, Novartis, Merck SD, Pfizer, Roche, Schering, and Schering Plough. Lee Lokey is an employee of the Network for Medical Communication and Research (NMCR) in Atlanta, Georgia. NMCR provides certain independent medical education and analytic services to Bristol-Myers Squibb for which fees are paid to NMCR. Lee Lokey does not directly receive any funds from Bristol-Myers Squibb.

Ancillary