Imatinib Mesylate for Philadelphia Chromosome-positive, Chronic-Phase Myeloid Leukemia after Failure of Interferon-α: Follow-Up Results¹ Free

Molecular abnormalities associated with the Philadelphia chromosome (Ph) have been causally implicated in the pathophysiology and development of Ph-positive leukemias (1, 2, 3). The Bcl-Abl fusion gene resulting from the translocation t(9;22)(q34;q11) can be produced through three different breakpoints in the BCR gene on chromosome 22. The three main resulting fusion proteins are p210, p190, and p230; other variants are rare (4, 5, 6, 7). Different forms of the Bcl-Abl gene and products are associated with different presentations of Ph-positive leukemias; p210 is associated with CML,⁴ p190 with Ph-positive acute lymphocytic leukemia, and p230 with either an indolent course of CML or chronic neutrophilic leukemia.

Recent therapies for CML have aimed at suppressing the Ph-positive cells so as to alter the natural course of the disease. Allogeneic SCT has resulted in long-term, event-free survival rates of 30–80% and 1-year mortality rates of 5–50%, depending on patient age, donor status (related versus unrelated), degree of matching, and other factors (e.g., cytomegalovirus status, the preparative regimen used, and prophylaxis for graft-versus-host disease or fungal infections; Refs. 8, 9, 10, 11, 12). Regimens containing IFN-α have produced CHR rates of 50–80% and cytogenetic response rates of 30–70%, with the latter classified as major (<35% Ph-positive cells) in 10–50% and complete (0% Ph-positive cells) in 5–30%. This variation in results arose from several factors, including disease-related risk factors, dose schedules of IFN-α, patient and physician compliance, patient selection, and the use of cytarabine with IFN-α. Median survival durations were 5–7 years for all patients and up to 9 years in patients considered to have “good-risk” disease. Achievement of a complete cytogenetic response has been associated with 10-year survival rates of 70–85% (13, 14, 15, 16, 17, 18, 19, 20, 21, 22). Other investigational treatment approaches have included the use of homoharringtonine, decitabine, intensive chemotherapy with autologous SCT, immunomodulation (with linomide or granulocyte-macrophage colony-stimulating factor), antiangiogenesis therapy, arsenic trioxide, farnesyl transferase inhibitors, and gene-targeted approaches (23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35).

Imatinib mesylate (STI571; Gleevec, Glivec) is a small molecular weight phenylaminopyrimidine designed to potently and selectively inhibit the Bcr-Abl tyrosine kinase in addition to other tyrosine kinases associated with c-kit and the platelet-derived growth factor receptor (36, 37). Imatinib mesylate was synthesized by identifying a lead compound in a high-throughput in vitro screen for tyrosine kinase inhibitors and then optimizing its activity against specific kinases. Imatinib mesylate acts specifically by inhibiting the binding site for ATP to the Abl kinase, thus blocking the phosphorylation of tyrosine on substrate protein. Preclinical studies showed that imatinib mesylate could suppress the leukemic cell growth of p210 and p190 cell lines as well as fresh samples from patients with CML. In Phase I–II studies, imatinib mesylate has shown encouraging results against CML in all three phases, chronic, accelerated, and blastic (38, 39, 40, 41, 42). Preliminary data have shown that in chronic-phase CML that is resistant to IFN-α therapy, imatinib mesylate therapy produced a CHR rate of 90% or above, major cytogenetic response rates of 60%, complete cytogenetic response rates of 40%, and 18-month estimated freedom from progression and survival rates of 89 and 95%, respectively (40). Until now, the multinational, multi-institutional nature of studies of imatinib mesylate, in which different laboratories have conducted the hematological, marrow, and cytogenetic studies and the data from different sources must be periodically transferred to a centralized database, has not allowed detailed outcomes analysis. Moreover, 15% of the patients enrolled in the studies have not been evaluable for various reasons.

Herein, we present long-term follow-up results for 261 patients with chronic-phase CML that did not respond to IFN-α therapy or were intolerant to IFN-α therapy who were treated with imatinib mesylate at our institution. The study objectives were to evaluate the incidence and durability of cytogenetic responses over a longer follow-up period and to analyze pretreatment and therapy-related factors associated with response and survival.

Patients in this analysis were treated in one of two Novartis-sponsored studies: a Phase II pivotal study of chronic-phase CML after IFN-α treatment failure (study 110, involving 149 patients previously included and analyzed as part of the multi-institutional multinational study; Ref. 40) and an expanded-access study (study 113, involving 112 patients) in which the eligibility criteria were expanded to include more patients until imatinib mesylate was approved by the Food and Drug Administration and hence became more available. Entry criteria were similar for both trials; eligible patients were 18 years or older and had adequate performance status (a score of 0–3 on an Eastern Cooperative Oncology Group scale), renal function (creatinine level of less than twice the upper limit of normal), hepatic function (bilirubin, aspartate aminotransferase, and alanine aminotransferase levels of less than twice the upper limit of normal), and cardiac function (those with New York Heart Association grade 3–4 cardiac disease were excluded). Patients could not have received hydroxyurea within the previous 7 days, IFN-α or cytarabine within the previous 14 days, or investigational agents within 28 days of starting imatinib mesylate. Women of child-bearing age were required to have a negative pregnancy test before starting imatinib mesylate, and all patients at risk were required to use barrier contraception during therapy. Patients provided written informed consent to participate before entry into the study; the study was reviewed and approved by the internal review board of the institution and performed in accordance with the Declaration of Helsinki.

Patients in the present study were required to have Ph-positive chronic-phase CML and had either failed to respond to IFN-α therapy at weekly doses of 25 MIU or more (according to hematological or cytogenetic response criteria) or could not tolerate IFN-α therapy.

Chronic-phase CML was defined as the presence in peripheral blood or bone marrow of a blast count of <15%, a basophil count of <20%, and a blast + promyelocyte count of <30% as well as a platelet count of more than 100 × 10⁹/liter. Hematological failure during IFN-α therapy was defined as the appearance of hematological resistance [failure to achieve CHR (as defined in the section on “Response” criteria below) after 6 or more months of IFN-α therapy] or relapse (disease recurrence after a CHR, with increases in WBC count to 20 × 10⁹/liter or more confirmed twice, 2 weeks apart, during IFN-α therapy). Concomitant hydroxyurea treatment was allowed for up to 50% of the treatment duration. Cytogenetic failure during IFN-α therapy was defined as resistance (the appearance of 65% or more Ph-positive cells in the bone marrow after >12 months of IFN-α therapy) or relapse (an increase in the proportion of Ph-positive cells of >30% documented on two occasions, or a single increase to 65% or more). Intolerance of IFN-α therapy was defined as the appearance of any nonhematological toxic effect of grade 3 or higher (according to the National Cancer Institute’s Common Toxicity Criteria) that persisted for >1 month in patients receiving IFN-α at a dose of 25 MIU/week or higher. The appearance of life-threatening toxic effects such as seizures, severe depression, immune-mediated complications, and severe neurotoxicity was considered evidence of intolerance, regardless of their duration.

Imatinib mesylate was given as an oral dose of 400 mg daily. Escalation of dose to 400 mg given twice daily was considered for patients who did not obtain a CHR after 3 or more months of therapy, those who relapsed after achieving CHR, or those who did not achieve a major cytogenetic response (<35% Ph-positive cells) after 12 months of therapy.

Dose reductions of imatinib mesylate for nonhematological or hematological toxic effects were performed as follows. For grade 2 persistent nonhematological toxic effects, therapy was interrupted until recovery to grade 1 or less and then resumed at the original dose level. If grade 2 effects reappeared, treatment was interrupted again until recovery and then resumed at a daily dose of 300 mg. For grade 3 or 4 nonhematological toxicity, therapy was interrupted until recovery to grade 2 or less and then resumed at a daily dose of 300 mg. For a grade 3–4 hematological effect (granulocyte count of <10⁹/liter or a platelet count of less than 50 × 10⁹/liter), therapy was interrupted until the effect resolved to grade 2 or better. If the toxic effect resolved within 2 weeks, treatment was resumed at the original dose of 400 mg daily. If grade 3–4 effects reappeared or if it had persisted for longer than 2 weeks, therapy was interrupted until the effect resolved to grade 2 or less and then resumed at a reduced daily dose of 300 mg. Patients developing anemia received transfusions of blood or blood products at the discretion of the investigator, or erythropoietin, 40,000 units given s.c. once a week until the hemoglobin level increased to 12 g/dl or more.

No concomitant anticancer drugs were administered. Patients with WBC counts of 20.0 × 10⁹/liter or more were given allopurinol 300 mg p.o. daily, which was discontinued after the WBC stabilized. Treatment with anagrelide (Agrylin) or leukapheresis (a maximum of one procedure/week) was permitted for up to 3 weeks.

Complete blood counts and serum chemistry evaluations were performed weekly during the first 12 weeks, every other week for the next 3 months, and every 6 weeks thereafter. Bone marrow studies, including morphological and cytogenetic or fluorescence in situ hybridization (iFISH) analysis, were performed every 3 months. Patients were followed for survival at least every 3 months. Drug safety parameters were evaluated at each visit and graded according to the National Cancer Institute Common Toxicity Criteria.

Response criteria have been described previously (13, 14, 15). A CHR was defined as a WBC count of less than 10 × 10⁹/liter, a platelet count of less than 450 × 10⁹/liter, no immature cells (blasts, promyelocytes, and myelocytes) in the peripheral blood, and disappearance of all signs and symptoms related to leukemia (including palpable splenomegaly) lasting for at least 4 weeks. Response was further categorized by the best cytogenetic response: complete if no Ph-positive cells were present, partial if the proportion of Ph-positive cells was between 1 and 34%, and minor if the proportion of Ph-positive cells declined to between 35 and 90%. Major cytogenetic response was defined as the sum of complete plus partial cytogenetic responses (i.e., all patients in whom Ph-positive cell proportions were <35%). The evaluation of cytogenetic response was judged by standard cytogenetic analysis of metaphase spreads, not by iFISH. Time to disease progression was calculated from the time the treatment began until the first reported appearance of accelerated- or blastic-phase disease, discontinuation of therapy because of unsatisfactory response, or death. Survival was calculated from the time the treatment began until death from any cause or last follow-up (43, 44).

Univariate and multivariate analyses were performed to identify potential prognostic factors and their association with major cytogenetic response rate and survival rate (45). The χ² test was used to identify prognostic factors, which were then included as terms in a multivariate regression model for response (46). Factors retaining significance in the multivariate model were interpreted as being independently predictive of major cytogenetic response. Because of the low number of events, no multivariate analysis of survival could be conducted.

A total of 261 patients with chronic-phase CML that failed to respond to IFN-α therapy participated in one of the two studies. Their characteristics are summarized in Table 1. When imatinib mesylate therapy was begun, of the 249 patients evaluable for response (Table 2), 142 patients had active disease and 107 patients were in CHR. The median duration of chronic phase was 33 months (range, 1–221 months). The overall median follow-up time was 17 months (range, 1–21 months); median follow-up for study 110 was 20 months (range, 12–21 months) and that for study 113 was 8.5 months (range, 1–14 months). Two hundred fifteen patients (82%) have been treated for at least 6 months, and 158 patients (61%) have been treated for at least 12 months. At the time of follow-up, 253 of the 261 patients (97%) were alive, and 241 (92%) were in chronic phase on imatinib mesylate therapy. Three patients have died after disease progression to accelerated (n = 1) or blastic (n = 2) phase; 5 patients died of other causes (1 from infection, 1 from cardiac events, and 3 from other causes).

Overall rates of hematological and cytogenetic responses to imatinib mesylate therapy are shown in Table 2. The CHR rate in our study group was 94% among patients with active disease. The cytogenetic response rate was 71%, with 62% of patients having achieved a major cytogenetic response and 45% a complete cytogenetic response. Response rates at 3, 6, and 12 months of therapy are shown in Table 3, demonstrating an increase in the cumulative major and complete cytogenetic response rates with continued therapy. Potential associations between pretreatment characteristics and achievement of a major cytogenetic response are shown in Table 4. Splenomegaly, leukocytosis, thrombocytosis, basophilia, any peripheral blasts or marrow blasts >5%, and clonal evolution were all inversely related to major cytogenetic response. Patients who started therapy within 12 months of diagnosis and those with <90% Ph-positive cells before therapy had significantly higher response rates. Patients with IFN-α intolerance or with hematological or cytogenetic relapse responded better than those with hematological or cytogenetic resistance.

In the multivariate analysis, we grouped patients into subsets according to some of the variables that had been found in univariate analysis to be associated with major cytogenetic response (e.g., splenomegaly present or absent); other factors were not included because of small sample size (e.g., hepatomegaly), apparent lack of association with response (e.g., sex), or close association with another similar variable (e.g., % marrow blasts versus % peripheral blasts). The results, summarized in Table 5, suggest that time from diagnosis to imatinib mesylate therapy, percentage of marrow basophils, percentage of Ph-positive cells before therapy, and prior response to IFN-α (hematological resistance versus others) were independently associated with achievement of major cytogenetic response. According to these results, patients could be assigned to one of into three prognostic risk groups: good, intermediate, or poor, with evident breakpoints probability of response (Pr; 0.8 or above, 0.5–0.7, and <0.4) and predicted major cytogenetic response rates of 93, 53, and 34%, respectively.

Patients in the good risk category (Pr > 0.8) included those with none or one of the four adverse factors (hematological resistance, diagnosis to therapy >12 months, marrow basophils >5%, and Ph-positive cells at start >90%). Those in the poor risk category (Pr <0.4) had three or four adverse features. Patients in the intermediate-risk category (Pr, 0.5–0.7) were those with two adverse features (Table 5).

Of the 261 patients treated, 241 (92%) remain alive in chronic phase and are still taking imatinib mesylate. The other 20 patients stopped therapy because of the development of hematological resistance (n = 4), transformation to accelerated or blastic phase (n = 9), toxic side effects (n = 2; liver function abnormalities), or other causes (1 intercurrent illness, 2 pregnancies, 1 sudden unrelated death, and 1 noncompliance). Eight patients have died, 7 after discontinuation of imatinib mesylate and 1 during imatinib mesylate therapy.

The estimated 18-month freedom from progression rate was 93% (Fig. 1), and that for survival was 96% (Fig. 2). Factors associated with worse survival in univariate analyses were leukocytosis (P = 0.05), elevated percentages of blast cells in the peripheral blood or marrow (P = 0.01), elevated percentage of basophils in the marrow (P = 0.04), and the presence of cytogenetic clonal evolution (P < 0.01). The latter appeared to have a significant impact on outcome difference, even at an early stage of follow-up. Because of the low rates of events in the survival analysis (only 20 patients stopped therapy and only 8 died), we could not evaluate possible predictors of survival by multivariate analysis.

Another important issue we considered is whether showing a response to imatinib mesylate after a short treatment period of 3–6 months could predict long-term outcome. This issue would be significant with regard to patient selection for continuation of imatinib mesylate versus referral for allogeneic SCT. We therefore conducted a landmark analysis of patients who were alive and still on imatinib mesylate therapy at 3 months and at 6 months in relation to their freedom from disease progression. The results are shown in Figs. 3 and 4.

Among the 243 patients who were alive and taking imatinib mesylate at 3 months, 131 (54%) had achieved a cytogenetic response (complete, partial, or minor). The estimated 18-month survival rates differed for those who had achieved a cytogenetic response (100%) and those who had not (95%; P < 0.001; Fig. 3,A). Similarly, the 18-month freedom from progression rate was better for patients who had achieved a cytogenetic response than for those who had not (100% versus 85%; P < 0.001; Fig. 3,B). Among the 205 patients who were alive and taking imatinib mesylate at 6 months, 96 (47%) had achieved a major cytogenetic response. The survival rate seemed to be better (Fig. 4,A; P = 0.07), and the freedom from progression rate (Fig. 4 B; P = 0.007) was better in patients who had achieved a major cytogenetic response than those who had not. However, the clinically relevant differences were too small at the current short follow-up to justify selection of patients for allogeneic SCT or other therapy according to whether they had achieved a particular cytogenetic response at 3 or 6 months of imatinib mesylate therapy.

Among the 155 patients who achieved a major cytogenetic response, 142 (92%) continue to have a durable cytogenetic response at the time of last follow-up, whereas 13 (8%) showed loss of major cytogenetic response (defined as an increase of Ph-positive cells by 30% or more documented on two occasions, or an increase to 65% or more documented once), after a median response duration of 6 months (range, 3–15 months). These 13 patients included 5 of 112 who had achieved a complete cytogenetic response, and 8 of 43 patients who had achieved a partial cytogenetic response.

Among the 22 patients who achieved a minor cytogenetic response, 2 (9%) have thus far shown loss of cytogenetic response (defined as an increase of Ph-positive cells to >90%) after a median of 6 and 7 months, respectively.

Another issue yet to be resolved is whether imatinib mesylate improves survival in patients with chronic-phase CML after IFN-α failure. Given the significant activity of imatinib mesylate, it is highly unlikely that any randomized comparative study of imatinib mesylate versus other “standards of care” would ever be conducted. We thus elected to compare the current results with those of previous studies of patients in whom IFN-α had failed who had been given other therapies, recognizing the preliminary nature of the comparison because of the short follow-up time on the imatinib mesylate therapy arm and the small number of events. Because the present study included a high proportion of patients with IFN-α intolerance (and such patients historically may have a better prognosis than those whose disease is refractory to IFN-α), we elected to compare our patients to those who had experienced hematological or cytogenetic resistance or relapse during IFN-α therapy and for whom therapy was stopped while the disease was in chronic phase. Patients who underwent allogeneic SCT in chronic phase were censored at the time of transplant. From 1988 through 2000, 131 patients were identified who met these criteria. Their median age was 44 years. Of these patients, 61 had experienced either hematological resistance (n = 25) or relapse (n = 36), and 70 had experienced either cytogenetic resistance (n = 45) or relapse (n = 25). Subsequent therapies were as follows: hydroxyurea (64 patients), homoharringtonine (41 patients), combination regimens (16 patients), and other (10 patients). The patients in the current study were older, but they had better performance status and lower incidences of hematological resistance or relapse and splenomegaly (Table 6; Refs. 47 and 48). Although the follow-up time is short, survival was significantly better among the imatinib-mesylate group than among the “other-treatments” comparison group (Fig. 5; P = 0.004). A multivariate analysis for survival conducted in the total study group identified older age, splenomegaly, and anemia to be independently significant predictors. Type of therapy (imatinib mesylate versus other) remained the most significant independent prognostic factor for survival, favoring imatinib mesylate therapy over other regimens (P = 0.004). However, only four events have thus far occurred in the imatinib mesylate arm, which limits the confidence in the multivariate analysis.

Side effects of imatinib mesylate in this study were similar to those reported previously for the entire imatinib mesylate treatment study. The severe (grade 3–4) toxic effects experienced are shown in Table 7. Only 2 patients had to discontinue imatinib mesylate therapy for unacceptable severe toxic effects; no patients died in chronic phase from imatinib mesylate-related complications.

Our study of patients with chronic-phase CML after failure of IFN-α therapy is unique in several aspects. It is the first large-scale, single-institution study evaluating the long-term results of patients treated uniformly with imatinib mesylate (400 mg/day given p.o.). It assessed pretreatment factors that might be associated with response and outcome after imatinib mesylate therapy, including some factors that were not included in the multinational studies, as well as other treatment-associated variables. Because of the large, well-defined historical database that exists at The University of Texas M. D. Anderson Cancer Center, we were able to compare the survival of patients in this study to that of patients whose disease had failed to respond to IFN-α (hematological or cytogenetic response) who had been treated with programs not including imatinib mesylate.

The longer-term follow-up results in this study of patients with chronic-phase CML after IFN-α failure treated with imatinib mesylate were very encouraging. Overall, 94% of patients achieved CHR, and 71% had a cytogenetic response, which was major in 62% and complete in 45%. The estimated 18-month freedom from progression and survival rates were 93 and 96%, respectively. Thus, the results with imatinib mesylate therapy continue to be positive with longer follow-up. The cytogenetic response rates also seem to be sustained and improving.

Better major cytogenetic response rates to imatinib mesylate therapy were observed in patients who had experienced relapse or had been intolerant to IFN-α therapy than in those treated for cytogenetic or hematological resistance. Similarly, better responses were noted in patients with less aggressive forms of disease (normal platelet counts, low blast and basophil levels, and no cytogenetic clonal evolution) and in patients with low tumor burden (absence of splenomegaly, in CHR, or with Ph suppression) or those treated within 12 months from diagnosis. These results suggest that frontline therapy with imatinib mesylate, either alone or in combination with other drugs, may further improve major and complete cytogenetic response rates; we are already observing rates of up to 70 and 50%, respectively, in our institutional studies of frontline imatinib mesylate for newly diagnosed CML (49). These surrogate endpoints have been associated previously with high long-term survival rates. Complete cytogenetic responses after IFN-α therapy have been associated with 10-year survival rates of as high as 80%. In this study, multivariate analysis identified a high percentage (>90%) of Ph-positive cells before therapy (P = 0.01), long interval between diagnosis a therapy (P = 0.02), marrow basophilia (P = 0.05), and hematological resistance to IFN-α (P = 0.06) as being independent factors predictive of poor cytogenetic response. Interestingly, older age, which has been associated previously with poor prognosis in CML, did not affect success with imatinib mesylate therapy. This finding is consistent with previous experience with other cancers (e.g., hairy cell leukemia) in which a new and highly effective therapy reduces or eliminates the significance of previously established prognostic factors. The prognostic model devised in this study allowed the identification of different risk groups with expected major cytogenetic relapse rates of 93, 53, and 34%, respectively. Patients in the poor-risk subgroup may in the future be offered investigational approaches such as imatinib mesylate combinations or other agents (e.g., homoharringtonine, decitabine, farnesyl transferase inhibitors, and arsenic trioxide) if they are not candidates for allogeneic SCT. The low number of events at this follow-up time precluded our analyzing factors that predict progression-free survival and survival rates. Of interest, cytogenetic clonal evolution, a “fingerprint” for clonal escape from the control of disease through Bcr-Abl suppression, was associated with a significantly worse outcome at this early phase of analysis (18-month survival rate of 84% versus 98% for others; P < 0.01).

Initial response to therapy with agents such as IFN-α has been included previously in the development of risk-oriented treatment strategies. In our previous IFN-α-containing regimens, patients who had not obtained a cytogenetic response after 12 or more months of IFN-α therapy were advised to consider allogeneic SCT when available. Combinations of pretreatment risk group and response to therapy have also been proposed for use in selecting patients for risk-oriented strategies (50). In view of the known curative potential of allogeneic SCT, albeit at a risk of mortality and significant morbidity, a reasonable question is whether imatinib mesylate therapy should be continued or whether allogeneic SCT should be considered in such a group, depending on the early response to imatinib mesylate therapy. Our analysis showed statistically significant, but clinically less relevant, differences in outcome according to whether patients had or had not achieved a cytogenetic response after 3 months or a major cytogenetic response after 6 months of imatinib mesylate therapy. Thus, at this follow-up time, we cannot rationally design risk-oriented approaches that consider early allogeneic SCT for patients who do not achieve a particular cytogenetic response early (at 3–6 months) in the course of imatinib mesylate therapy.

Previous experience has led to estimated yearly mortality rates of 10–15% for patients in chronic-phase CML that had failed to respond or had relapsed during therapy (47, 48). Because of the extremely positive experience with imatinib mesylate therapy, it may not be possible to design randomized comparative trials of imatinib mesylate versus “standard of care” for patients who do not respond to IFN-α therapy to assess the relative survival benefit conferred by imatinib mesylate. We thus compared the results of the study reported here to results in similar patients whose disease had become unresponsive to IFN-α therapy but was treated with other approaches. Our findings suggest superiority of imatinib mesylate over other regimens, with estimated 18-month survival rates of 96% versus 78% (Fig. 5; P < 0.01). Imatinib mesylate therapy is now the new standard of care for patients with chronic-phase CML and IFN-α failure. Combinations of imatinib mesylate with IFN-α, cytarabine, homoharringtonine, decitabine, or other compounds hopefully will further improve the complete cytogenetic and molecular response rates and thus the long-term prognosis for patients with this disease.