Background: The first BCR-ABL1 tyrosine kinase inhibitor (TKI), imatinibmesylate (IM), has become a first-line therapy for chronic phase (CP) chronic myeloid leukemia (CML). However, approximately one third of IM-treated patients discontinue therapy due to an inadequate response or adverse event. More potent second or third generation TKIs such as nilotinib, dasatinib, radotinib, bosutinib and ponatinib have developed and these agents have shown high rates of hematologic and cytogenetic responses after failure of IM therapy. Although the new European Leukemia Net (ELN) recommendation serves provisionally the definitions of the response to second-line therapy, early molecular milestones which are associated with the best long-term are not confirmed. The aim of this study was to identify 3-month molecular milestone for predicting long-term survival to second-line therapy using second or third generation TKIs in CP CML patients who showed a failure or warning to IM.

Methods: Between March 2005 and January 2014, 198 CP CML patients with failure or warning to IM (defined by 2013 ELN recommendation) had been treated with nilotinib, dasatinib, radotinib, bosutinib or ponatinib as a second-line therapy. Among them, 180 patients had available molecular data at 3 months from the initiation of second-line therapy. Based on receiver operating characteristic (ROC) curveanalysis, the predictive cutoffs of BCL-ABL1 transcripts at 3 months for progression-free survival (PFS), and overall survival (OS) were evaluated. OS included any death regardless of causes, and PFS included progression to AP or BP as well as death resulting from any reason. OS and PFS were also collected on patients who were treated with other TKIs after failure of second-line TKI therapy.

Results: A total of 180 patients were treated with second-line TKI, dasatinib (n=66), nilotinib (n=61), radotinib (n=44), bosutinib (n=7), and ponatinib (n=2). 119 men and 61 women were included and their median age was 42 years (range, 15-75). Using a ROC curve analysis, BCR-ABL1 transcript level 7.93% on the international scale, at 3 months were predictive cutoffs for both PFS and OS. The median follow-up for survivors since the start of second-line TKI was 78.73 months (range, 6.3-114.0 months). 104 patients continue on therapy and 76 patients were permanently discontinued due to intolerance (n=38), failure (n=20), progression (n=14), and others (n=4). The7-year PFS and OS were 82.6% and 85.3%, respectively. The patients with transcript levels below 7.93% at 3 months had significantly better 7-year PFS (95.1% vs. 60.4%; P < 0.001) and OS (96.3% vs. 67.9%; P < 0.001). After adjusting for potential factors affecting PFS and OS in univariate analyses, multivariate analyses showed that BCR-ABL1 transcript of 7.93% at 3 months was the independent predictor for PFS (RR of 8.37, P < 0.001) and OS (RR of 13.53, P = 0.001). In addition, increasing age (RR of 1.05, P = 0.023) and presence of BCR-ABL1 kinase domain abnormalities (KDA) at baseline (RR of 5.83, P = 0.007) were associated with a lower OS.

Conclusions: Our data showed BCR-ABL1 transcript of 7.93% at 3 months was an early independent predictor for long-term survival to second-line TKIs in IM-resistant CP CML. It implies that the patients who failed an achievement of reduction of BCR-ABL1transcripts to this level may require more precise monitoring on second-line therapy, allowing early clinical intervention using other third-line TKI therapy or allografting.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.