Page 100 - 2019_08-Haematologica-web

C. Fava et al.
ation, 211 patients (72%) were on treatment with ima- tinib and 82 patients (28%) with either nilotinib (n=58), dasatinib (n=23), or bosutinib (n=1). There were no differ- ences in age, sex, Sokal score and type of transcript between imatinib and second generation TKI. One hun- dred and sixty-two patients (55%) discontinued in first line, 117 patients (40%) in second line, 13 patients (4.5%) in third line, and one patient in fourth line. Among those who discontinued imatinib, 73 patients (35%) had been pre-treated with α-interferon (IFN) and seven patients had been submitted to allogeneic stem cell transplantation. Median duration of treatment with any TKI was 87 months (IQR 59-117) for all patients, 96 months (IQR 62- 120) for imatinib patients, and 73 (IQR 51-98) months for second generation TKI patients (P=0.002). Median dura- tion of treatment with the last TKI was 77 months (IQR 54-111) for all patients, and 50 months (IQR 32; 66) for second generation TKI patients. Median duration of DMR was 46 months (IQR 30-73) for all patients, 53 months (IQR 33-82) for imatinib patients, and 36 months (IQR 25- 46) for second generation TKI patients (P<0.001). Overall, all patients but one had an optimal early response to last treatment. At three months of last TKI, 34% of patients were in MMR, 40% were in CCyR and/or had a transcript ≤1%, and 25% were in PCyR and/or had a transcript ≤10%.
At treatment discontinuation the response was as fol- lows: undetectable transcript in 16% of patients, MR4 in 35% of patients, MR4.5 in 31% of patients, and MR5 in 18%. There was no difference in the grade of molecular response at discontinuation between patients on imatinib and patients on second generation TKI (P=0.315).
Relapses and treatment-free remission
At 12 months, the estimated TFR was 69% (95%CI: 64- 75) for all patients (Figure 1A), 68% (95%CI: 62-74) for imatinib patients (Figure 1B), 73% (95%CI: 64-83) for sec- ond generation TKI patients (Figure 1C).
Median follow up was 34 months (IQR 24-53) for all patients, 42 months (IQR 26-56) for imatinib patients, and 26 months (IQR 21-34) for second generation TKI patients. At median follow up, TFR was 62% (95%CI: 56- 68) for all patients (at 34 months), 60% (95%CI: 54-67) for imatinib patients (at 42 months), 67% (95%CI: 57-78) for second generation TKI patients (at 26 months) (Figure 1). There was no significant difference in TFR between patients who had discontinued imatinib first-line versus imatinib after IFN versus further lines (P=0.35), and there was no difference in TFR between patients who discontin- ued second generation TKI frontline (n=33) versus second- line for intolerance (n=30) versus second-line for resistance (n=16) (P=0.16).
Overall, 114 patients (39%) resumed treatment. Reasons for resuming were: loss of MR4 (19%), loss of MMR (70%), loss of CCyR (9%), other (2%). The reasons for restarting imatinib and second generation TKI were similar (P=0.13). Overall median time to restart treatment was six months (IQR 4-11). Although 75% of patients had restarted treatment by the end of the first year, the last treatment resumption was after 105 months of TFR. Median time to loss of MR4 was three months (IQR 2-7); median time to loss of MMR was four months (IQR 3-7), and median time to loss of CCyR was five months (IQR 4-6). Median time from loss of response to restarting treat- ment was one month (IQR 0-2).
AB
C
Figure 1. Kaplan-Meier curves for Italian patients who discontinued tyrosine kinase inhibitor (TKI). (A) Overall population. (B) Patients who discontinued imatinib. (C) Patients who discontinued second generation TKI. Estimated treatment-free remission (TFR) is reported at 12 months for the overall population; at 12, 26 (medi- an follow up for patients who discontinued second generation TKI), and 42 (median follow up for patients who discontinued imatinib) months for imatinib; at 12 and 26 months (median follow up for patients who discontinued second generation TKI) for second generation TKI. N: number.
1592
haematologica | 2019; 104(8)