reatment. In the IFN arm, withdrawal of consent (14%) was the most frequent reason for discontinuation of first-line therapy, and the most frequent reason for cross over to the Gleevec arm was severe intolerance to treatment (26%) and progression (14%).
The primary efficacy endpoint of the study was progression-free survival (PFS). Progression was defined as any of the following events:progression to accelerated phase or blast crisis (AP/BC), death, loss of CHR or MCyR, or in patients not achieving a CHR an increasing WBC despite appropriate therapeutic management. The protocol specified that the progression analysis would compare the intent to treat (ITT) population:patients randomized to receive Gleevec were compared with patients randomized to receive IFN. Patients that crossed over prior to progression were not censored at the time of cross-over, and events that occurred in these patients following cross-over were attributed to the original randomized treatment. The estimated rate of progression-free survival at 84 months in the ITT population was 81.2 % [95% CI: 78, 85] in the Gleevec arm and 60.6 % [56, 65] in the IFN arm (p<0.0001, log-rank test), (Figure 1). With 7 years follow up there were 93 (16.8%) progression events in the Gleevec arm:37(6.7%) progression to AP/BC, 31(5.6%) loss of MCyR, 15 (2.7%) loss of CHR or increase in WBC and 10 (1.8% ) CML unrelated deaths. In contrast, there were 165 (29.8% ) events in the IFN+Ara-C arm of which 130 occurred during first-line treatment with IFN-Ara-C. The estimated rate of patients free of progression to accelerated phase (AP) or blast crisis (BC) at 84 months was 92.5 %[90, 95] in the Gleevec arm compared to the 85.1%, [82, 89] (p≤0.001) in the IFN arm, (Figure 2). The annual rates of any progression events have decreased with time on therapy.The probability of remaining progression free at 60 months was 95% for patients who were in complete cytogenetic response (CCyR) with molecular response (≥3 log reduction in Bcr-Abl transcripts as measured by quantitative reverse transcriptase polymerase chain reaction) at 12 months, compared to 89% for patients in complete cytogenetic response but without a major molecular response and 70% in patients who were not in complete cytogenetic response at this time point (p<0.001).
Figure 1 Progression Free Survival (ITT Principle)
Figure 2 Time to Progression to AP or BC (ITT Principle)
A total of 71 (12.8%) and 85 (15.4%) patients died in the Gleevec and IFN+Ara-C group, respectively. At 84 months the estimated overall survival is 86.4% (83, 90) vs. 83.3% (80, 87) in the randomized Gleevec and the IFN+Ara-C group, respectively (p=0.073 log-rank test). The hazard ratio is 0.750 with 95% CI 0.547 – 1.028. This time-to-event endpoint may be affected by the high crossover rate from IFN+Ara-C to Gleevec.Major cytogenetic response, hematologic response, eva luation of minimal residual disease (molecular response), time to accelerated phase or blast crisis and survival were main secondary endpoints. Response data are shown in Table 14. Complete hematologic response, major cytogenetic response and complete cytogenetic response were also statistically significantly higher in the Gleevec arm compared to the IFN + Ara-C arm (no cross-over data considered for eva luation of responses). Median time to CCyR in the 454 responders was 6 months (range 2-64 months, 25th to 75th percentiles = 3 to 11 months) with 10% of responses seen only after 22 months of therapy).
Ta
