Responses to second-line tyrosine kinase inhibitors are durable: an intention-to-treat analysis in chronic myeloid leukemia patients

Abstract

Second-generation tyrosine kinase inhibitors (2G-TKIs) are effective at inducing complete cytogenetic responses (CCyRs) in approximately half of chronic myeloid leukemia patients treated while still in the chronic phase and after failing imatinib. It is less clear whether these responses are durable. In the present study, we report the clinical outcome of 119 patients who received a 2G-TKI as second-line treatment while still in the chronic phase. In an intention-to-treat analysis, the 4-year probabilities of overall and event-free survival were 81.9% and 35.3%, respectively. Sixty-two patients discontinued the initial 2G-TKI because of resistance or intolerance. To further explore the durability of cytogenetic responses, irrespective of the need for a third-line TKI, we used the concept of "current CCyR-survival" (c-CCyRS). The c-CCyRS at 4 years was 54.4%. After introduction of a 2G-TKI, 77 patients had a 3-month BCR-ABL1/ABL1 transcript ratio of ≤ 10% and had significantly superior overall survival (91.3% vs 72.1%, P = .02), event-free survival (49.3% vs 13.0%, P < .001), and c-CCyRS (67.2% vs 11.2%, P = .0001) compared with the 33 patients with ratios > 10%. The 3-month molecular response was the only independent predictor for overall survival. Using an intention-to-treat analysis, we have shown that the responses to second-line therapies are durable. Patients destined to fare poorly can be identified early during therapy.

Figure 1

OS, EFS, CI of CCyR, and c-CCyRS according to the molecular response to first-choice 2G-TKI at 3 months. The 77 patients who after 3 months of second-line therapy had BCR-ABL1/ABL1 ratios of ≤ 10% (solid line) had a significantly superior OS (A; 91.3% vs 72.1%, P = .02), EFS (B; 49.3% vs 13.0%, P < .001), and CI of CCyR (C; 76% vs 4%, P < .001) compared with the 30 patients with transcript ratios > 10% (broken line). Whiskers represent 95% confidence intervals. Ten patients had missing quantitative real-time PCR data and 2 were already receiving third-line TKI therapy and therefore were not included in the analysis. The patients with a transcript ratio ≤ 10% had a significantly higher probability of c-CCyRS (D; defined as the probability of being alive and in CCyR at a given time point): 67.2% for patients with a transcript level ≤ 10% (blue line) and 11.2% (P = .001) for patients with a higher transcript level (red line).

Figure 2

OS, EFS, and c-CCyRS in the whole population and according to the cytogenetic response at 12 months. (A) Probabilities of OS (gray solid line), EFS (black broken line), and c-CCyRS (black solid line) in the whole population. The 4-year probability of OS, EFS, and c-CCyRS were 81.9%, 35.3%, and 54.4%, respectively. The figure shows how the EFS continues to decline over time, whereas the c-CCyRS plateaued at approximately 55% after the first year of therapy, reflecting that the responses are stable once they have been achieved. (B) EFS (broken line) and c-CCyRS (solid line) according to the cytogenetic response achieved at 12 months. The 48 patients who were in CCyR at 12 months, represented in black, had a 4-year probability of EFS of 71.8%; conversely, these patients had an excellent 4-year probability of c-CCyRS of 90%. In contrast, the 51 patients who were not in CCyR at 12 months, represented in gray, had a EFS of 5.6% (P < .001) and a c-CCyRS of 7.5% (P = .008)

Figure 3

Evolution in the BCR-ABL1 transcript level on first-choice second-line 2G-TKI according to the molecular response achieved after 3 months on second-line therapy. Patients who had a BCR-ABL1/ABL1 ratio > 10% after 3 months of therapy (red circles) had a minimal reduction in the transcript level during subsequent follow-up, remaining always above the 1% mark (equivalent to a 2 log reduction commonly identified with CCyR). Patients with a transcript level at 3 months of ≤ 10% are represented by blue circles. The top row represents patients who discontinued their first-choice 2G-TKI at the various time points according to their transcript levels at 3 months.

Figure 4

c-CCyRS according to the dose intensity received by the patients during the first 3 months of therapy and reason for dose adjustment. The 39 patients (gray line) who had their therapy with the first-choice 2G-TKI temporarily discontinued or dose reduced because of hematologic toxicity during the first 3 months of treatment had a significantly worse 4-year probability of c-CCyRS than the 54 patients who did not require dose modification, represented by the black line, (26.4% vs 64.5%, P = .001). Conversely, the 23 patients who had their dose of 2G-TKI modified because of nonhematologic toxicity (broken black line) had a 4-year c-CCyRS comparable to that of patients with good tolerance to therapy (55.5% vs 64.5%, P = .64), which was significantly higher than that of patients with hematological intolerance (P = .02).