In a large phase III trial (IRIS) comparing imatinib versus IFN/Ara-c, imatinib demonstrated significant higher rate of major (MCyR) and complete cytogenetic response (CCyR) and improved progression free survival (PFS). However a large number of pts (n= 354) assigned to the IFN/Ara-c arm crossed-over to imatinib therapy after a median of 9 months. Thus, we performed a retrospective analysis comparing outcome of pts assigned to imatinib in the IRIS trial (

NEJM 2003;348: 994
) and pts assigned to IFN/Ara-c arm included in the French multicenter (CML 91) trial before imatinib became available (
NEJM 1997; 337: 223
).

Methods: Pts in the IRIS trial received imatinib 400 mg daily. Pts in the CML 91 trial received IFN alpha 2b 5 MIU/m2 daily plus monthly courses of Ara-c 20 mg/m2 for 10 days. Inclusion criteria of both studies were similar: pts with Ph+ CP CML diagnosed within 6 months before randomization; hydroxyurea or anagrelide were allowed if needed. We compared pts who actually received the assigned study treatment (n=551 for IRIS and n=325 for CML91).

Results: We analyzed the results with a cut-off period of 42 months, median follow-up for alive pts being also 42 months in both groups. The imatinib treated population included more pts who were male, with high or unknown Sokal score and higher % of blasts, and with lower % of basophils at diagnosis than IFN plus Ara-c population. Evaluation included MCyR, CCyR, PFS (i.e. evolution to accelerated phase/blast crisis or death: to be consistent with the CML91 design), and overall survival. Results are summarized in the table below.

Best response (CCyR) was observed in 81% (IRIS) and 32% (CML91) of pts.

At the time of analysis 7% of all pts had never achieved confirmed MCyR but were still under study treatment in the IRIS; 8% were off treatment. These were 26% and 23% in the CML 91. Within all Sokal risk groups, imatinib was superior to IFN/Ara-c in terms of MCR, CCyR, PFS and overall survival. Using a landmark analysis, for pts who achieved MCyR at 1 year (n = 437 for IRIS and n = 125 for CML91), the survival at 36 months was similar in both groups: 95% for IRIS and 96% for CML 91.

Conclusion: This historical comparison shows that for first line therapy for newly diagnosed CP CML, imatinib is superior to prolonged therapy with IFN/Ara-c for rate of MCyR, CCyR, PFS and overall survival. Pts who do not have access to imatinib may have long PFS and overall survival with IFN/Ara-c therapy.

IRIS studyCML 91 studylogrank test**
*95% CI ; ** by 42 months 
Estimated probability of MCR at 12 months 85 (82–88)* 39 (33–45) 
Estimated probability of MCR at 36 months 93 (91–96) 61 (55–68) p<0.0001 
Estimated probability of CCyR at 12 months 70 (66–74) 14 (10–18) 
Estimated probability of CCyR at 36 months 87 (83–90) 42 (35–48) p<0.0001 
Progression free survival at 12 months 98 (97–99) 96 (94–98) 
Progression free survival at 36 months 90 (87–93) 82 (77–87) p = 0.004 
Overall survival at 12 months 99 (98–100) 98 (96–100) 
Overall survival at 36 months 92 (89–94) 84 (80–88) p<0.0001 
IRIS studyCML 91 studylogrank test**
*95% CI ; ** by 42 months 
Estimated probability of MCR at 12 months 85 (82–88)* 39 (33–45) 
Estimated probability of MCR at 36 months 93 (91–96) 61 (55–68) p<0.0001 
Estimated probability of CCyR at 12 months 70 (66–74) 14 (10–18) 
Estimated probability of CCyR at 36 months 87 (83–90) 42 (35–48) p<0.0001 
Progression free survival at 12 months 98 (97–99) 96 (94–98) 
Progression free survival at 36 months 90 (87–93) 82 (77–87) p = 0.004 
Overall survival at 12 months 99 (98–100) 98 (96–100) 
Overall survival at 36 months 92 (89–94) 84 (80–88) p<0.0001 

Author notes

Corresponding author

Sign in via your Institution