Very long-term follow-up results of imatinib mesylate therapy in chronic phase chronic myeloid leukemia after failure of interferon alpha therapy


  • Hagop Kantarjian and Jenny Shan designed research, performed research, analyzed data, and wrote the article. Susan O'Brien, Guillermo Garcia-Manero, and Jorge Cortes designed research, performed research, analyzed data, and wrote the article. Stefan Faderl, Farhad Ravandi, and Elias Jabbour performed research, analyzed data, and wrote the article.



The long-term outcome of patients with chronic phase chronic myeloid leukemia treated with imatinib after failure of interferon alpha therapy has not been detailed.


In total, 368 patients were analyzed. Univariate and multivariate survival analyses were conducted using standard statistical methods.


Overall, 247 patients (67%) achieved a complete cytogenetic response (CCyR). Of the 327 patients who were studied, 207 patients (63%) achieved a major molecular response (MMR), and 99 patients (30%) had undetectable breakpoint cluster region/c-abl oncogene (BCR-ABL) levels at some time during therapy. The estimated 10-year survival rate was 68%, the progression-free survival rate was 67%, and the event-free survival rate was 51%. In multivariate analysis, age ≥60 years, hemoglobin <10 g/dL, bone marrow basophils ≥5%, any peripheral blasts, and clonal evolution were independent adverse factors for survival. The estimated 7-year survival rate according to the presence of no factors (n = 154), 1 or 2 factors (n = 190), or ≥3 factors (n = 24) were 93%, 70%, and 25%, respectively (P < .01). Achieving an MMR, a CCyR, or a partial cytogenetic response at 12 months was associated with significantly better 10-year survival rate in a landmark analysis (10-year survival rate, 80%-90%) compared with achieving a minor cytogenetic response or a complete hematologic response (10-year survival rate, 55%-65%) or another response (10-year survival rate, 10%). In a landmark analysis that included imatinib response at 12 months, achieving a major cytogenetic response or better (hazard ratio, 0.12; P < .001) and achieving a complete hematologic response or a minor cytogenetic response (hazard ratio, 0.36; P = .003) were significant favorable prognostic factors.


The current results indicated that the estimated 10-year survival rate of 68% for patients with chronic myeloid leukemia who receive imatinib after failure on interferon has improved. Cancer 2012;118: 3116–22. © 2011 American Cancer Society.


Imatinib mesylate, a breakpoint cluster region/c-Abl oncogene (Bcr-Abl) selective tyrosine kinase inhibitor (TKI), has revolutionized the treatment and prognosis of patients with Philadelphia-chromosome (Ph)-positive chronic myeloid leukemia (CML). Imatinib is now the frontline standard of care in CML.1-4 With imatinib therapy, a complete cytogenetic response (CCyR) is obtained in 80% to 85% of patients and is durable at 5 years in 67% of patients. The estimated 8-year survival rate of patients is 84% to 93%, depending on whether non-CML deaths are included or censored.2, 4

The incidence of CML in the United States is about 5000 new cases annually.5 With a median survival of 3 to 6 years before the widespread use of imatinib, the prevalence of CML was estimated at approximately 15,000 to 30,000 cases. These patients either would have undergone allogeneic stem cell transplantation (SCT) or would have received interferon alpha-based therapy. Thus, it is estimated that approximately 15,000 to 20,000 patients in the United States may have received imatinib after interferon. The short-term follow-up of these patients revealed favorable results6-8; however, it is important to determine their long-term prognosis to decide whether such therapy continues to be beneficial and to identify which patients should be considered for alternative therapies, including second-generation TKIs and allogeneic SCT. This analysis focuses on the very long-term results of patients with chronic phase CML who received imatinib after interferon failure at our institution. Compared with multi-institutional or pharmaceutical-sponsored studies, single-institutional studies may be better at following the course of patients for events or progression after they are taken off the protocol imatinib regimen.


Adults with Ph-positive CML in chronic phase who were treated at our institution with imatinib therapy after interferon failure were included in this analysis. These patients were entered on institutional review board-approved protocols that were available at our institution during the period. Informed consent was obtained according to institutional guidelines and in compliance with the Declaration of Helsinki. The details of eligibility criteria, treatment plan, and modifications have been detailed previously.6-8

Response criteria were as previously described. A cytogenetic response (CyR) was categorized as complete (CCyR) if Ph-positive metaphases were reduced to 0%, partial (PCyR) if Ph-positive metaphases were reduced to 1% to 35%, and minor if Ph-positive metaphases were reduced to 36% to 90%.6-8 A major CyR (MCyR) included CCyRs and PCyRs. Survival was calculated from the start of therapy until death from any cause. The time to progression to accelerated or blastic phase (progression-free survival [PFS]) was calculated from the start of therapy until the development of accelerated or blastic phase or until death from any cause on or off TKI therapy. Event-free survival (EFS) was calculated from the start of therapy until the development of any event leading to the patient being off the study, including loss of MCyR, loss of hematologic response, unacceptable toxicity, death from any cause, or transformation to accelerated or blastic phase. Survival, PFS, and EFS curves were analyzed with the Kaplan-Meier method and were compared by using log-rank tests.9 The cumulative incidence rates for CCyR, complete molecular response (CMR), and major molecular response (MMR) were estimated by taking into account the competing risks for events. Standard statistical methods were used for univariate and multivariate survival analyses.10

Follow-up studies included bone marrow studies, and cytogenetic and molecular analyses were performed every 3 months in the first year, every 6 months in years 2 to 5, and at least every 6 to 12 months thereafter. It is noteworthy that 95 patients were not followed regularly at our institution after deciding not to continue on the imatinib study protocol. Their subsequent outcome is followed only for death as an event. Therefore, they are censored at the time of last follow-up at M. D. Anderson Cancer Center unless they die, at which time they would be updated as death event for survival, PFS, and EFS.


In total, we analyzed 368 patients with Ph-positive CML in chronic phase who received imatinib therapy after failure on interferon alpha. Their characteristics are detailed in Table 1. Their median age was 54 years, and 121 patients (33%) were aged ≥60 years. The median follow-up of patients on study was 114 months (range, 1-132 months).

Table 1. Characteristics of the Study Group (n=368)
CharacteristicNo. of Patients (%)Median [Range]
  1. Abbreviations: BM, bone marrow; Ph+, Philadelphia chromosome-positive; WBC, white blood cells.

Age ≥60 yrs121 (33)54 [21-82]
Splenomegaly52/366 (14) 
Hemoglobin ≤12 g/dL149 (41)12.3 [7.3-16.9]
WBC >50 ×109/L38 (10)9.1 [1.8-240.5]
Platelets >450 ×109/L72 (20)241 [81-1116]
BM blasts ≥5%33 (9) 
Peripheral blasts present67 (18) 
BM basophils ≥5%44 (12) 
Peripheral basophils ≥7%36 (10) 
Cytogenetic clonal evolution66/368 (18) 
Duration of chronic phase, mo  
 <1236 (10) 
 12-35163 (44) 
 ≥36169 (46) 
Ph+ pretreatment <90%94 (26) 
Interferon alpha failure  
 Hematologic47 (13) 
 Cytogenetic171 (46) 
 Intolerance148 (40) 

Overall, 247 patients (67%) achieved a CCyR as their best cytogenetic response. Of 326 patients who had molecular studies, 207 (63%) achieved an MMR (BCR-ABL ratio on the international scale, ≤0.1%), and 99 (30%) had undetectable levels at some time during therapy. The cumulative incidence of MMR by 7 years was 55% (95% confidence interval [CI], 45%-64%), and the cumulative incidence of undetectable BCR levels was 35% (95% CI, 29%-42%). Considering the total denominator of 368 patients, in an intent-to-treat analysis at 7 years, 129 patients (35%) were in CCyR, 116 patients (32%) were in MMR, and 40 patients (11%) had undetectable BCR-ABL levels.

At the most recent follow-up, the estimated 10-year survival rate was 68%, the PFS rate was 67%, and the EFS rate was 51% (Fig. 1). Currently, 104 patients (28%) have died. The causes of death are indicated in Figure 2. Death in transformation was caused by accelerated phase disease complications in 13 patients and by blastic phase disease in 39 patients. Among the latter group, blastic phase was lymphoid in 9 patients, myeloid in 24 patients, and mixed lineage in 1 patient (5 patients had unknown blastic phase morphology).

Figure 1.

Survival, progression-free survival (Progression), and event-free survival (Event) are illustrated.

Figure 2.

Patient status is illustrated. CCyR indicates complete cytogenetic response; PCyR, partial cytogenetic response; MMR, major molecular response; CML, chronic myeloid leukemia.

Among 256 patients (70%) in whom imatinib therapy was discontinued, 11 died on imatinib therapy from non-CML causes (Fig. 2). One hundred eighteen patients (35%) stopped the study imatinib for reasons other than imatinib resistance, including no further follow-up at our institution (n = 95), patient request (n = 4), physician decision (n = 3), socioeconomic factors (n = 5), noncompliance (n = 3), and other malignancies or diseases (n = 8). One hundred twenty-seven patients failed imatinib, including 86 patients who remain in chronic phase and 41 patients who have transformed (blastic phase; n = 14; accelerated phase; n = 27) (Fig. 2).

Sixty-two of 127 patients (49%) who failed on imatinib therapy received salvage therapy with second-generation TKIs as first salvage (nilotinib, 14 patients; dasatinib, 34 patients; others, 14 patients), and 27 patients received second salvage therapy with second-generation TKIs. The estimated 5-year survival of patients from the time of second-generation TKIs as salvage therapy is 60%.

In total, 10 patients were referred for allogeneic SCT. Referral to transplantation was after the development of imatinib resistance in chronic phase for 4 patients and in transformation for 6 patients. Currently, 3 patients (30%) are alive without evidence of disease. (1 patient relapsed received dasatinib treatment).

Prognostic factors associated with survival are listed in Table 2. A multivariate analysis identified older age (≥60 years), severe anemia (hemoglobin <10 g/dL), bone marrow basophils ≥5%, the presence of any peripheral blasts, and clonal evolution at the start of imatinib therapy as independent adverse factors. Overall survival according to the number of adverse factors is illustrated in Figure 3 and Table 3. Three risk groups were generated to predict for long-term survival.

Figure 3.

Survival is illustrated according to the number of pretreatment adverse prognostic factors.

Table 2. Prognostic Factors Associated With Survival
  Survival Rate, %   
CategoryNo. of Patients7 Years10 YearsUnivariate PHRMultivariate P
  1. Abbreviations: BM, bone marrow; CML, chronic myeloid leukemia; HR, hazard ratio; IFN, interferon alpha; NS, nonsignificant; PB, peripheral blood; Ph+ mets, Philadelphia-positive metaphases; WBC, white blood cells.

Age, y      
CML duration, mo      
 <12367369.18 NS
 No3148073<.001 NS
Hemoglobin, g/dL      
Platelets, ×109/L      
 ≤4502968071.001 NS
WBC, ×109/L      
 <101978377<.001 NS
PB blasts, %      
BM blasts, %      
 <53357869.03 NS
PB basophils, %      
 <73327869.11 NS
BM basophils, %      
Clonal evolution      
Ph+ mets, %      
 ≤901118979.002 NS
IFN failure      
 Hematologic475956<.001 NS
Table 3. Survival According to the Number of Adverse Factorsa
   Survival Rate (95% CI), % 
Risk GroupNo. of Adverse Risk FactorsTotal No./ No. of Deaths7 Years10 YearsMedian Survival [Range], mo
  • Abbreviations: CI, confidence interval; NR, not reached.

  • a

    Adverse factors were age ≥60 years, hemoglobin <10 g/dL, bone marrow basophils ≥5%, any peripheral blasts, and the presence of cytogenetic clonal evolution.

Low riskNone154/1593 (88-97)89 (83-95)NR
Intermediate risk1-2190/7070 (63-77)58 (51-67)NR
High risk≥324/1925 (13-50)21 (10-45)39 [26-74]

The outcome of patients according to their 12-month response to imatinib is illustrated in Figure 4. Although patients who achieved an MMR, a CCyR, or a PCyR by 12 months had better outcomes, those who achieved even a minor CyR or even a complete hematologic response had better outcomes than nonresponders.

Figure 4.

Survival is illustrated according to 12-month response to imatinib (landmark analysis at 12 months [12mos]). mmr indicates major molecular response, CCyR, complete cytogenetic response; PCyR, partial cytogenetic response; mincyr, minor cytogenetic response; CHR, complete cytogenetic response.

By using a landmark analysis to include imatinib response by 12 months, we identified the achievement of an MCyR (hazard ratio, 0.12; P < .001), and a complete hematologic response or a minor CyR (hazard ratio, 0.36; P = .003) as significant prognostic factors for long-term outcome after the analysis was adjusted for covariates.


This study presents the unique experience of very long-term follow-up results from a single institution in patients who received imatinib after interferon failure. We noted favorable results with a 10-year survival rate of 68% and an EFS rate of 51%.

Several recent studies have analyzed the outcome of patients with CML on imatinib therapy in different contexts, and some have included patients who received imatinib after interferon failure.11, 12 In the study by Gambacorti-Passerini et al,11 832 patients who received imatinib were reviewed. Among them, 478 patients received imatinib as second-line therapy. Those authors focused on the long-term side effects and safety profile of imatinib, emphasizing the need for independent, multicenter assessment of outcomes in patients with CML. In a population-based study of patients with CML who were diagnosed in Sweden from 1973 to 2008, Bjorkholm et al demonstrated the impact of introducing imatinib therapy to a large CML population on long-term survival.12 In their analysis, the 5-year cumulative relative survival rates increased 20% to 50% during the years 1973 to 2000 up to 80% since 2001, the year when imatinib therapy started having increased penetration into CML therapy in Sweden. Those authors did not separate outcome according to patients who received imatinib as frontline therapy versus salvage therapy; however, they did demonstrate the overall impact of introducing imatinib therapy to the total CML population.

Our results and those of others8, 11, 12 are encouraging and suggest a major change in the natural course of CML after interferon failure, which, of course, is partly expected from the results of frontline imatinib therapy. This clarifies the favorable outcome of patients after interferon failure, a group whose median survival was 3 years when imatinib therapy was not available.7 In the current study, the estimated 10-year survival rate of patients was 68% compared with 20% to 30% in the historic group of patients who had failure on interferon therapy and had no access to imatinib.7 This suggests that most patients can be managed safely and effectively with imatinib after interferon failure and do not have to consider the option of allogeneic SCT as long as they are in CyR. Even if they demonstrate imatinib resistance in chronic phase, subsequent salvage therapy with second-generation TKIs (second salvage in the context of this analysis; interferon therapy frontline, imatinib therapy as salvage) can be safely entertained as a durable therapeutic approach as long as the patients are sensitive and responsive to imatinib therapy.

In multi-institutional studies, patients on TKI therapy often may not be followed adequately for events that occur after they are taken off the study TKI, except for death.1, 2 This is because of limited resources as well as difficulties in following patients (for example, with bone marrow studies or even routine blood tests that are not prespecified in the protocol requirements). Our single-institutional study offers the opportunity of more precise estimations of events after failure on TKI therapy, but it is still limited by a lack of detailed follow-up in 95 of 368 patients (26%). The variability in the definitions of EFS and PFS and its impact on perceived differences in outcome has been highlighted recently.13, 14 In our study, we used the most conservative definition of EFS in which all events are counted, including the event of taking a patient off TKI therapy for any reason. Therefore, a 10-year EFS rate of 51% represents a reasonably favorable outcome for this study group.

In summary, the current analysis confirms the efficacy and safety of continued imatinib therapy in patients with Ph-positive CML after interferon failure and reassures patients who are currently on such therapy in the oncology community, and their treating physicians, that continuation of this therapy is reasonable. Undertaking allogeneic SCT should be considered only in patients who develop imatinib resistance.


This study was supported by grant P01 CA049639 (The Therapy of CML) from the National Institutes of Health. Also supported by the Betty Foster Leukemia Research Fund.


Hagop Kantarjian and Jorge Cortes are consultants for Novartis and have research funding from Novartis, BMS and Pfizer. Elias Jabbour is a speaker for Novartis.